Novel Architecture for 3D model in virtual communities from detected face
This addresses the need for efficient avatar creation in virtual communities, but appears incremental as it builds on existing face detection and modification techniques.
The paper tackles the problem of creating personalized avatars from facial images by extracting, modifying, and characterizing faces, achieving satisfactory performance for practical use when tested on hundreds of images under uncontrolled conditions.
In this research paper we suggest how to extract a face from an image, modify it, characterize it in terms of high-level properties, and apply it to the creation of a personalized avatar. In this research work we tested, we implemented the algorithm on several hundred facial images, including many taken under uncontrolled acquisition conditions, and found to exhibit satisfactory performance for immediate practical use.